fluentd 1.20.0-aarch64-mingw-ucrt

data/lib/fluent/plugin/filter_record_transformer.rb

@@ -0,0 +1,325 @@
+#
+# Fluentd
+#
+#    Licensed under the Apache License, Version 2.0 (the "License");
+#    you may not use this file except in compliance with the License.
+#    You may obtain a copy of the License at
+#
+#        http://www.apache.org/licenses/LICENSE-2.0
+#
+#    Unless required by applicable law or agreed to in writing, software
+#    distributed under the License is distributed on an "AS IS" BASIS,
+#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#    See the License for the specific language governing permissions and
+#    limitations under the License.
+#
+
+require 'socket'
+require 'json'
+require 'ostruct'
+
+require 'fluent/env'
+require 'fluent/plugin/filter'
+require 'fluent/config/error'
+require 'fluent/event'
+require 'fluent/time'
+
+module Fluent::Plugin
+  class RecordTransformerFilter < Fluent::Plugin::Filter
+    Fluent::Plugin.register_filter('record_transformer', self)
+
+    helpers :record_accessor
+
+    desc 'A comma-delimited list of keys to delete.'
+    config_param :remove_keys, :array, default: nil
+    desc 'A comma-delimited list of keys to keep.'
+    config_param :keep_keys, :array, default: nil
+    desc 'Create new Hash to transform incoming data'
+    config_param :renew_record, :bool, default: false
+    desc 'Specify field name of the record to overwrite the time of events. Its value must be unix time.'
+    config_param :renew_time_key, :string, default: nil
+    desc 'When set to true, the full Ruby syntax is enabled in the ${...} expression.'
+    config_param :enable_ruby, :bool, default: false
+    desc 'Use original value type.'
+    config_param :auto_typecast, :bool, default: true
+
+    def configure(conf)
+      super
+
+      map = {}
+      # <record></record> directive
+      conf.elements.select { |element| element.name == 'record' }.each do |element|
+        element.each_pair do |k, v|
+          element.has_key?(k) # to suppress unread configuration warning
+          map[k] = parse_value(v)
+        end
+      end
+
+      if @keep_keys
+        raise Fluent::ConfigError, "`renew_record` must be true to use `keep_keys`" unless @renew_record
+      end
+
+      @key_deleters = if @remove_keys
+                        @remove_keys.map { |k| record_accessor_create(k) }
+                      end
+
+      placeholder_expander_params = {
+        log: log,
+        auto_typecast: @auto_typecast,
+      }
+      @placeholder_expander =
+        if @enable_ruby
+          # require utilities which would be used in ruby placeholders
+          require 'pathname'
+          require 'uri'
+          require 'cgi/escape'
+          RubyPlaceholderExpander.new(placeholder_expander_params)
+        else
+          PlaceholderExpander.new(placeholder_expander_params)
+        end
+      @map = @placeholder_expander.preprocess_map(map)
+
+      @hostname = Socket.gethostname
+    end
+
+    def filter_stream(tag, es)
+      new_es = Fluent::MultiEventStream.new
+      tag_parts = tag.split('.')
+      tag_prefix = tag_prefix(tag_parts)
+      tag_suffix = tag_suffix(tag_parts)
+      placeholder_values = {
+        'tag'        => tag,
+        'tag_parts'  => tag_parts,
+        'tag_prefix' => tag_prefix,
+        'tag_suffix' => tag_suffix,
+        'hostname'   => @hostname,
+      }
+      es.each do |time, record|
+        begin
+          placeholder_values['time'] = @placeholder_expander.time_value(time)
+          placeholder_values['record'] = record
+
+          new_record = reform(record, placeholder_values)
+          if @renew_time_key && new_record.has_key?(@renew_time_key)
+            time = Fluent::EventTime.from_time(Time.at(new_record[@renew_time_key].to_f))
+          end
+          @key_deleters.each { |deleter| deleter.delete(new_record) } if @key_deleters
+          new_es.add(time, new_record)
+        rescue => e
+          router.emit_error_event(tag, time, record, e)
+          log.debug { "map:#{@map} record:#{record} placeholder_values:#{placeholder_values}" }
+        end
+      end
+      new_es
+    end
+
+    private
+
+    def parse_value(value_str)
+      if value_str.start_with?('{', '[')
+        JSON.parse(value_str, **Fluent::DEFAULT_JSON_PARSE_OPTIONS)
+      else
+        value_str
+      end
+    rescue => e
+      log.warn "failed to parse #{value_str} as json. Assuming #{value_str} is a string", error: e
+      value_str # emit as string
+    end
+
+    def reform(record, placeholder_values)
+      placeholders = @placeholder_expander.prepare_placeholders(placeholder_values)
+
+      new_record = @renew_record ? {} : record.dup
+      @keep_keys.each do |k|
+        new_record[k] = record[k] if record.has_key?(k)
+      end if @keep_keys && @renew_record
+      new_record.merge!(expand_placeholders(@map, placeholders))
+
+      new_record
+    end
+
+    def expand_placeholders(value, placeholders)
+      if value.is_a?(String)
+        new_value = @placeholder_expander.expand(value, placeholders)
+      elsif value.is_a?(Hash)
+        new_value = {}
+        value.each_pair do |k, v|
+          new_key = @placeholder_expander.expand(k, placeholders, true)
+          new_value[new_key] = expand_placeholders(v, placeholders)
+        end
+      elsif value.is_a?(Array)
+        new_value = []
+        value.each_with_index do |v, i|
+          new_value[i] = expand_placeholders(v, placeholders)
+        end
+      else
+        new_value = value
+      end
+      new_value
+    end
+
+    def tag_prefix(tag_parts)
+      return [] if tag_parts.empty?
+      tag_prefix = [tag_parts.first]
+      1.upto(tag_parts.size-1).each do |i|
+        tag_prefix[i] = "#{tag_prefix[i-1]}.#{tag_parts[i]}"
+      end
+      tag_prefix
+    end
+
+    def tag_suffix(tag_parts)
+      return [] if tag_parts.empty?
+      rev_tag_parts = tag_parts.reverse
+      rev_tag_suffix = [rev_tag_parts.first]
+      1.upto(tag_parts.size-1).each do |i|
+        rev_tag_suffix[i] = "#{rev_tag_parts[i]}.#{rev_tag_suffix[i-1]}"
+      end
+      rev_tag_suffix.reverse!
+    end
+
+    # THIS CLASS MUST BE THREAD-SAFE
+    class PlaceholderExpander
+      attr_reader :placeholders, :log
+
+      def initialize(params)
+        @log = params[:log]
+        @auto_typecast = params[:auto_typecast]
+      end
+
+      def time_value(time)
+        Time.at(time).to_s
+      end
+
+      def preprocess_map(value, force_stringify = false)
+        value
+      end
+
+      def prepare_placeholders(placeholder_values)
+        placeholders = {}
+
+        placeholder_values.each do |key, value|
+          if value.kind_of?(Array) # tag_parts, etc
+            size = value.size
+            value.each_with_index do |v, idx|
+              placeholders.store("${#{key}[#{idx}]}", v)
+              placeholders.store("${#{key}[#{idx-size}]}", v) # support [-1]
+            end
+          elsif value.kind_of?(Hash) # record, etc
+            value.each do |k, v|
+              placeholders.store(%Q[${#{key}["#{k}"]}], v) # record["foo"]
+            end
+          else # string, integer, float, and others?
+            placeholders.store("${#{key}}", value)
+          end
+        end
+
+        placeholders
+      end
+
+      # Expand string with placeholders
+      #
+      # @param [String] str
+      # @param [Boolean] force_stringify the value must be string, used for hash key
+      def expand(str, placeholders, force_stringify = false)
+        if @auto_typecast && !force_stringify
+          single_placeholder_matched = str.match(/\A(\${[^}]+}|__[A-Z_]+__)\z/)
+          if single_placeholder_matched
+            log_if_unknown_placeholder($1, placeholders)
+            return placeholders[single_placeholder_matched[1]]
+          end
+        end
+        str.gsub(/(\${[^}]+}|__[A-Z_]+__)/) {
+          log_if_unknown_placeholder($1, placeholders)
+          placeholders[$1]
+        }
+      end
+
+      private
+
+      def log_if_unknown_placeholder(placeholder, placeholders)
+        unless placeholders.include?(placeholder)
+          log.warn "unknown placeholder `#{placeholder}` found"
+        end
+      end
+    end
+
+    # THIS CLASS MUST BE THREAD-SAFE
+    class RubyPlaceholderExpander
+      attr_reader :log
+
+      def initialize(params)
+        @log = params[:log]
+        @auto_typecast = params[:auto_typecast]
+        @cleanroom_expander = CleanroomExpander.new
+      end
+
+      def time_value(time)
+        Time.at(time)
+      end
+
+      # Preprocess record map to convert into ruby string expansion
+      #
+      # @param [Hash|String|Array] value record map config
+      # @param [Boolean] force_stringify the value must be string, used for hash key
+      def preprocess_map(value, force_stringify = false)
+        new_value = nil
+        if value.is_a?(String)
+          if @auto_typecast && !force_stringify
+            num_placeholders = value.scan('${').size
+            if num_placeholders == 1 && value.start_with?('${') && value.end_with?('}')
+              new_value = value[2..-2] # ${..} => ..
+            end
+          end
+          unless new_value
+            new_value = "%Q[#{value.gsub('${', '#{')}]" # xx${..}xx => %Q[xx#{..}xx]
+          end
+        elsif value.is_a?(Hash)
+          new_value = {}
+          value.each_pair do |k, v|
+            new_value[preprocess_map(k, true)] = preprocess_map(v)
+          end
+        elsif value.is_a?(Array)
+          new_value = []
+          value.each_with_index do |v, i|
+            new_value[i] = preprocess_map(v)
+          end
+        else
+          new_value = value
+        end
+        new_value
+      end
+
+      def prepare_placeholders(placeholder_values)
+        placeholder_values
+      end
+
+      # Expand string with placeholders
+      #
+      # @param [String] str
+      def expand(str, placeholders, force_stringify = false)
+        @cleanroom_expander.expand(
+          str,
+          placeholders['tag'],
+          placeholders['time'],
+          placeholders['record'],
+          placeholders['tag_parts'],
+          placeholders['tag_prefix'],
+          placeholders['tag_suffix'],
+          placeholders['hostname'],
+        )
+      rescue => e
+        raise "failed to expand `#{str}` : error = #{e}"
+      end
+
+      class CleanroomExpander
+        def expand(__str_to_eval__, tag, time, record, tag_parts, tag_prefix, tag_suffix, hostname)
+          instance_eval(__str_to_eval__)
+        end
+
+        (Object.instance_methods).each do |m|
+          undef_method m unless /^__|respond_to_missing\?|object_id|public_methods|instance_eval|method_missing|define_singleton_method|respond_to\?|new_ostruct_member|^class$/.match?(m.to_s)
+        end
+      end
+    end
+  end
+end

data/lib/fluent/plugin/filter_stdout.rb

@@ -0,0 +1,53 @@
+#
+# Fluentd
+#
+#    Licensed under the Apache License, Version 2.0 (the "License");
+#    you may not use this file except in compliance with the License.
+#    You may obtain a copy of the License at
+#
+#        http://www.apache.org/licenses/LICENSE-2.0
+#
+#    Unless required by applicable law or agreed to in writing, software
+#    distributed under the License is distributed on an "AS IS" BASIS,
+#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#    See the License for the specific language governing permissions and
+#    limitations under the License.
+#
+
+require 'fluent/plugin/filter'
+
+module Fluent::Plugin
+  class StdoutFilter < Filter
+    Fluent::Plugin.register_filter('stdout', self)
+
+    helpers :formatter, :compat_parameters, :inject
+
+    DEFAULT_FORMAT_TYPE = 'stdout'
+
+    config_section :format do
+      config_set_default :@type, DEFAULT_FORMAT_TYPE
+    end
+
+    # for tests
+    attr_reader :formatter
+
+    def configure(conf)
+      compat_parameters_convert(conf, :inject, :formatter)
+      super
+      @formatter = formatter_create
+    end
+
+    def filter_stream(tag, es)
+      es.each { |time, record|
+        begin
+          r = inject_values_to_record(tag, time, record)
+          log.write @formatter.format(tag, time, r)
+        rescue => e
+          router.emit_error_event(tag, time, record, e)
+        end
+      }
+      log.flush
+      es
+    end
+  end
+end

data/lib/fluent/plugin/formatter.rb

@@ -0,0 +1,75 @@
+#
+# Fluentd
+#
+#    Licensed under the Apache License, Version 2.0 (the "License");
+#    you may not use this file except in compliance with the License.
+#    You may obtain a copy of the License at
+#
+#        http://www.apache.org/licenses/LICENSE-2.0
+#
+#    Unless required by applicable law or agreed to in writing, software
+#    distributed under the License is distributed on an "AS IS" BASIS,
+#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#    See the License for the specific language governing permissions and
+#    limitations under the License.
+#
+
+require 'fluent/env'
+require 'fluent/plugin/base'
+require 'fluent/plugin/owned_by_mixin'
+require 'fluent/time'
+
+module Fluent
+  module Plugin
+    class Formatter < Base
+      include OwnedByMixin
+      include TimeMixin::Formatter
+
+      configured_in :format
+
+      PARSER_TYPES = [:text_per_line, :text, :binary]
+      def formatter_type
+        :text_per_line
+      end
+
+      def format(tag, time, record)
+        raise NotImplementedError, "Implement this method in child class"
+      end
+    end
+
+    class ProcWrappedFormatter < Formatter
+      def initialize(proc)
+        super()
+        @proc = proc
+      end
+
+      def format(tag, time, record)
+        @proc.call(tag, time, record)
+      end
+    end
+
+    module Newline
+      module Mixin
+        include Fluent::Configurable
+
+        DEFAULT_NEWLINE = if Fluent.windows?
+                            :crlf
+                          else
+                            :lf
+                          end
+
+        config_param :newline, :enum, list: [:lf, :crlf], default: DEFAULT_NEWLINE
+
+        def configure(conf)
+          super
+          @newline = case newline
+                     when :lf
+                       "\n".freeze
+                     when :crlf
+                       "\r\n".freeze
+                     end
+        end
+      end
+    end
+  end
+end

data/lib/fluent/plugin/formatter_csv.rb

@@ -0,0 +1,92 @@
+#
+# Fluentd
+#
+#    Licensed under the Apache License, Version 2.0 (the "License");
+#    you may not use this file except in compliance with the License.
+#    You may obtain a copy of the License at
+#
+#        http://www.apache.org/licenses/LICENSE-2.0
+#
+#    Unless required by applicable law or agreed to in writing, software
+#    distributed under the License is distributed on an "AS IS" BASIS,
+#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#    See the License for the specific language governing permissions and
+#    limitations under the License.
+#
+
+require 'fluent/plugin_helper'
+require 'fluent/plugin/formatter'
+require 'csv'
+
+module Fluent
+  module Plugin
+    class CsvFormatter < Formatter
+      Plugin.register_formatter('csv', self)
+
+      include PluginHelper::Mixin
+      helpers :record_accessor
+
+      config_param :delimiter, default: ',' do |val|
+        ['\t', 'TAB'].include?(val) ? "\t".freeze : val.freeze
+      end
+      config_param :force_quotes, :bool, default: true
+      # "array" looks good for type of :fields, but this implementation removes tailing comma
+      # TODO: Is it needed to support tailing comma?
+      config_param :fields, :array, value_type: :string
+      config_param :add_newline, :bool, default: true
+
+      def csv_cacheable?
+        !!owner
+      end
+
+      def csv_thread_key
+        csv_cacheable? ? "#{owner.plugin_id}_csv_formatter_#{@usage}_csv" : nil
+      end
+
+      def csv_for_thread
+        if csv_cacheable?
+          Thread.current[csv_thread_key] ||= CSV.new("".force_encoding(Encoding::ASCII_8BIT), **@generate_opts)
+        else
+          CSV.new("".force_encoding(Encoding::ASCII_8BIT), **@generate_opts)
+        end
+      end
+
+      def configure(conf)
+        super
+
+        @fields = fields.select{|f| !f.empty? }
+        raise ConfigError, "empty value is specified in fields parameter" if @fields.empty?
+
+        if @fields.any? { |f| record_accessor_nested?(f) }
+          @accessors = @fields.map { |f| record_accessor_create(f) }
+          mformat = method(:format_with_nested_fields)
+          singleton_class.module_eval do
+            define_method(:format, mformat)
+          end
+        end
+
+        @generate_opts = {col_sep: @delimiter, force_quotes: @force_quotes, headers: @fields,
+                          row_sep: @add_newline ? :auto : "".force_encoding(Encoding::ASCII_8BIT)}
+      end
+
+      def format(tag, time, record)
+        csv = csv_for_thread
+        line = (csv << record).string.dup
+        # Need manual cleanup because CSV writer doesn't provide such method.
+        csv.rewind
+        csv.truncate(0)
+        line
+      end
+
+      def format_with_nested_fields(tag, time, record)
+        csv = csv_for_thread
+        values = @accessors.map { |a| a.call(record) }
+        line = (csv << values).string.dup
+        # Need manual cleanup because CSV writer doesn't provide such method.
+        csv.rewind
+        csv.truncate(0)
+        line
+      end
+    end
+  end
+end

data/lib/fluent/plugin/formatter_hash.rb

@@ -0,0 +1,35 @@
+#
+# Fluentd
+#
+#    Licensed under the Apache License, Version 2.0 (the "License");
+#    you may not use this file except in compliance with the License.
+#    You may obtain a copy of the License at
+#
+#        http://www.apache.org/licenses/LICENSE-2.0
+#
+#    Unless required by applicable law or agreed to in writing, software
+#    distributed under the License is distributed on an "AS IS" BASIS,
+#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#    See the License for the specific language governing permissions and
+#    limitations under the License.
+#
+
+require 'fluent/plugin/formatter'
+
+module Fluent
+  module Plugin
+    class HashFormatter < Formatter
+      include Fluent::Plugin::Newline::Mixin
+
+      Plugin.register_formatter('hash', self)
+
+      config_param :add_newline, :bool, default: true
+
+      def format(tag, time, record)
+        line = record.to_s
+        line << @newline if @add_newline
+        line
+      end
+    end
+  end
+end

data/lib/fluent/plugin/formatter_json.rb

@@ -0,0 +1,62 @@
+#
+# Fluentd
+#
+#    Licensed under the Apache License, Version 2.0 (the "License");
+#    you may not use this file except in compliance with the License.
+#    You may obtain a copy of the License at
+#
+#        http://www.apache.org/licenses/LICENSE-2.0
+#
+#    Unless required by applicable law or agreed to in writing, software
+#    distributed under the License is distributed on an "AS IS" BASIS,
+#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#    See the License for the specific language governing permissions and
+#    limitations under the License.
+#
+
+require 'fluent/plugin/formatter'
+require 'fluent/oj_options'
+
+module Fluent
+  module Plugin
+    class JSONFormatter < Formatter
+      include Fluent::Plugin::Newline::Mixin
+
+      Plugin.register_formatter('json', self)
+
+      config_param :json_parser, :string, default: 'oj'
+      config_param :add_newline, :bool, default: true
+
+      def configure(conf)
+        super
+
+        if @json_parser == 'oj'
+          if Fluent::OjOptions.available?
+            @dump_proc = Oj.method(:dump)
+          else
+            log.info "Oj isn't installed, fallback to JSON as json parser"
+            @dump_proc = JSON.method(:generate)
+          end
+        else
+          @dump_proc = JSON.method(:generate)
+        end
+
+        # format json is used on various highload environment, so re-define method to skip if check
+        unless @add_newline
+          define_singleton_method(:format, method(:format_without_nl))
+        end
+      end
+
+      def format(tag, time, record)
+        json_str = @dump_proc.call(record)
+        "#{json_str}#{@newline}"
+      ensure
+        json_str&.clear
+      end
+
+      def format_without_nl(tag, time, record)
+        @dump_proc.call(record)
+      end
+    end
+  end
+end

data/lib/fluent/plugin/formatter_ltsv.rb

@@ -0,0 +1,44 @@
+#
+# Fluentd
+#
+#    Licensed under the Apache License, Version 2.0 (the "License");
+#    you may not use this file except in compliance with the License.
+#    You may obtain a copy of the License at
+#
+#        http://www.apache.org/licenses/LICENSE-2.0
+#
+#    Unless required by applicable law or agreed to in writing, software
+#    distributed under the License is distributed on an "AS IS" BASIS,
+#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#    See the License for the specific language governing permissions and
+#    limitations under the License.
+#
+
+require 'fluent/plugin/formatter'
+
+module Fluent
+  module Plugin
+    class LabeledTSVFormatter < Formatter
+      include Fluent::Plugin::Newline::Mixin
+
+      Plugin.register_formatter('ltsv', self)
+
+      # http://ltsv.org/
+
+      config_param :delimiter, :string, default: "\t".freeze
+      config_param :label_delimiter, :string, default: ":".freeze
+      config_param :replacement, :string, default: " ".freeze
+      config_param :add_newline, :bool, default: true
+
+      def format(tag, time, record)
+        formatted = ""
+        record.each do |label, value|
+          formatted << @delimiter if formatted.length.nonzero?
+          formatted << "#{label}#{@label_delimiter}#{value.to_s.gsub(@delimiter, @replacement)}"
+        end
+        formatted << @newline if @add_newline
+        formatted
+      end
+    end
+  end
+end